Monitoring canid scent marking in space and time using a biologging and machine learning approach

Latrines are important sites for intraspecific olfactory communication in mammals, especially for solitary or widely distributed species. Communal latrines give visitors access to information about other visitors, notably conspecific chemical cues, even in their absence. Chemical communication has evolved to allow information transfer among individuals that, due to other ecological constraints, do not co-occur in time. Latrines can be difficult to find and monitor but provide useful information about the behavioral ecology of otters. The aim of this study was to describe the behaviors of Neotropical otters (Lontra longicaudis, Olfers, 1818) in two communal latrines using video-camera traps. A total of 1,651 one-minute footage of otters visiting the latrines were used to elaborate an ethogram that included individuals (1) passing by, (2) rubbing, (3) scent marking, (4) scratching, (5) in vigilance, (6) smelling, (7) defecating, (8) urinating, (9) digging, (10) self-grooming, and (11) interacting with others. These results suggest that latrines are not only used by Neotropical otters to deposit feces and urine but that they also play a role in intraspecific communication. We suggest that L. longicaudis latrines function as information centers where individuals can monitor the location and activities of potential sexual partners and/or competitors.

Scent over‐marking occurs when an animal deposits its scent mark on top of the scent mark of a conspecific. Over‐marking may provide advantages in the transfer of information to the individual whose scent is on top but not to the individual whose scent is on the bottom. We tested the hypothesis that over‐marking is a competitive form of olfactory communication and that male prairie voles would over‐mark the scent marks of same‐sex conspecifics more than those of same‐sex siblings. Two age‐matched male voles (first male and second male) were placed successively into an arena in which they were allowed to explore freely and scent mark for 15 min at age 12, 20, 28, 36, 44, and 52 d. The first male was placed into a clean arena, whereas the second male was placed into an arena containing either the scent marks of an age‐matched male sibling or nonsibling. Age affected the total number of scent marks deposited by the voles; 12‐20‐d‐old voles deposited fewer scent marks, over‐marks and adjacent marks than did 28‐52‐d‐old voles. Sibship did not affect the total number of scent marks deposited by the first and second voles but did affect the number of over‐marks and adjacent marks deposited by the second vole. Siblings received significantly fewer over‐marks and adjacent marks than did nonsiblings; this effect was most dramatic after the voles reached 28 d of age, a time coincident with the onset of puberty. Males separated from siblings and housed singly at 44‐d‐old and tested at 52‐d‐old, deposited significantly more over‐marks and adjacent marks in arenas if the first vole was a nonsibling than if it was a sibling. This differential scent‐marking supports the hypothesis that over‐marking and adjacent marking are used as competitive forms of olfactory communication by male prairie voles.

We investigated the effect of sex, social status and life-stage on three behaviours (pasting, dry-pasting and pawing) associated with olfactory communication in the spotted hyena (Crocuta crocuta). Pasting is a behaviour used to place a fatty, anal scent gland secretion on vegetation. Dry-pasting behaviour resembles pasting except that scent is not deposited, but instead scent pasted on vegetation by other clan members is smeared on the ventral side of a dry-pasting animal. Pawing is vigorous scratching of the ground with one then the other front paw, which deposits scent from inter-digital glands at scratching sites. Our focus was olfactory communication within social groups termed clans. We obtained data from individually known members of three large clans in the Serengeti National Park, in northern Tanzania. Data were collected at clan communal dens, which are important social centres, and elsewhere in the clan territory. Adult female’s pasting and dry-pasting rates significantly declined with increasing age which indicated that younger adult females invested more than older females in depositing scent and anointing themselves with scent from other clan members. Females pawed the ground rarely. Pasting and dry-pasting rates of reproductively active males (mostly immigrant and a few natal males) above median social status were significantly higher than rates of those below median status at the clan communal den, but not elsewhere in the clan territory. This may result from a reduced opportunity for males below median status to scent mark at the clan den because of the presence of socially dominant males. Alternatively, a lower-ranking male (i.e. more recent immigrant) may gain fitness benefits from concealing its olfactory identity from den-bound juvenile females, because this would increase the number of females that would consider it a potential mate throughout its tenure in the clan. Males above median status pawed the ground at a higher rate both at the communal den and elsewhere in the territory than males below median status. We conclude that these sex differences reflect the importance of (1) olfactory communication and scent mixing between females during their initial years of social integration into the adult hierarchy; (2) male-male competition in terms of pawing throughout the territory and possibly pasting at the den; (3) social communication and scent mixing between males mostly away from the communal den.

Scent marking posture frequency of intact dogs and bitches was observed at two locations (kennel and open yard) in pre estrous and estrous phase in the domestic dog and effect of breed, age, parity and litter size of the gestation length. This study was carried out in Security and Guarding Dog Training Center, Cairo, Egypt. Twenty individuals (10 dogs and 10 bitches) for observation of scent marking posture, licking, sniffing the urine of female, tonguing response, sniffing and licking female perineal region, in pre estrous and estrous phase. Also data of 100 bitches were collected from the stud book record through the period between 2013-2015 analyzed to study the effect of breed, age, parity and litter size on the gestation length. The results indicated that there were a significant difference (p ≤ 0.05) between breeds, litter size and the gestation length, while no significant difference of age and parity of bitches. Also, there was a significant increase in mean frequencies of high leg raising, slight leg raising, squatting squat raising and raised leg display in dogs than in bitches and no significant difference in mean frequency of stand squatting. A significant increase in mean frequencies of licking and sniffing the urine of a bitch, tonguing response and sniffing and licking female perineal region and male investigating in the open yard.

Karen McComb

The behaviour of marking trees by the brown bear occurs throughout the range of presence of the species. It has recently been recognised as a mean of intra-specific chemical communication, besides the likely function of ecto-parasite removal, and evidence from grizzlies showed that scent marking of trees is mainly performed by adult males during the breeding season. However, detailed studies on this behaviour in the Euroasian brown bear are lacking. We conducted a three year study on a wild bear population of 50 individuals in the Eastern Italian Alps, using camera traps. We aimed to assess the use of trees by bears’ age and sex classes, its temporal variation, and to determine key habitat and human disturbance drivers of the intensity of use of rub trees. Camera trapping yielded more than 500 videos of bears from more than 9000 camera trapping days, age and sex classes were identified in 59% of the videos. Results showed that rubbing was mainly performed by adult males, with females and sub-adults that only occasionally rubbed. Rubbing was performed more during the breeding season, confirming that the main function of this behaviour is associated to males' breeding strategy. Olfactory investigation was performed by bears of all ages and sexes, indicating the importance of scent marking at rub trees for intra-specific communication. We used the camera trapping event rate as a raw index of intensity of usage of rub trees and found it to be affected by aspect, type of roads and passage of motor vehicles at the sites, while the passage of people did not affect it. We also estimated bear occupancy and detectability and found that the latter was influenced by trail type and distance from roads and buildings. Our study provides a first contribution to the use of rub trees by brown bears in the Alps. Download the complete issue.

The study of vocal behavior can reveal important aspects of how and why a species communicates in relation to ecological and social challenges. We here focus on vocal communication in golden-backed uakaris (Cacajao melanocephalus), diurnal, pitheciine monkeys that exhibit fission-fusion social organization and typically inhabit dense forests that limit the potential for visual communication. Moreover, the species spends little time engaged in tactile or olfactory communication, e.g., social grooming and scent marking, respectively. Hence, vocalizations may be very important for the coordination of social organization in these monkeys. We 1) categorized golden-backed uakari vocalizations, 2) ascertained their behavioral context, and 3) investigated whether golden-backed uakari calls can encode information about the signaler. We observed the monkeys during 2 wet seasons in the flooded igapo forest of Jau National Park, Brazil. We showed that golden-backed uakaris have 9 call types in their vocal repertoire, all distinguishable by ear and from analysis of spectrograms. Some calls, e.g., play-specific calls, were used only in particular behavioral contexts, and by individuals of specific age, whereas others were emitted under a range of situations. The structure of the loud tcho call varied among individuals, and according to behavioral context, i.e., whether individuals were foraging/feeding, traveling, or performing agonistic interactions. This knowledge of the species’ vocal repertoire is valuable for surveying the monkeys acoustically in habitats where visual surveys are difficult.

Despite the growing interest in olfactory communication, few studies have examined neural mechanisms controlling scent marking. Compared to efforts to understand how odor cues affect recipient animals, little attention has focused on any aspect of signal sending. Interest in the signaling animal is often limited to identifying the chemical structure of the odor involved and to determining how and where it is produced. Yet to understand olfactory communication processes, we must also study the mechanisms that determine when and where animals actively signal by releasing or depositing odors.

Background: Adolescent depression is highly prevalent in low- and middle-income countries (LMICs). Identifying top key risk factors is necessary to inform effective prevention program design. Machine learning (ML) offers a powerful approach to analyze complex multidomain of data to identify the most relevant predictors and estimate risks for mental health problems. This paper applies ML to study risks for adolescent depression to enhance adolescent depression prevention efforts in LMICs. Methods: Six ML approaches (e.g., Explainable Boosting Machine, random forests, and XGBoost) were applied to study the risks of depression. Data were drawn from a digital health intervention study conducted in Kenya (year 2024–2025, n = 269). Multiple domains of childhood and adolescent adversity and stress experiences were used to predict adolescent depression (using PHQ9-A). Findings: We found that ML was a valuable approach in the early identification of adolescents at risk for depression. Among the six ML approaches applied, the random forest approach outperformed other ML approaches, especially when multiple domains of risks were included. We also found that childhood adversity or home adversity alone were not strong predictors for depression. Adding adolescent stress experiences and community school adversity experiences significantly improves the accuracy and predictability of depression. Using the top-15 and top-20 ranking factors, we achieved 74.8% and 75.1% accuracy in depression prediction, which was similar to the accuracy when all 49 adverse/stress factors were included in the predictive model (78.3%). Conclusions: Innovative ML and modern predictive modeling approaches have the potential to transform modern preventive mental health care by better utilizing multidomain data to identify individuals at risk for developing depression early and identify top risk factors (for targeted individuals and/or populations). Findings from ML can inform tailored intervention design to better mitigate risks in order to prevent depression problem development. They can also inform the better utilization of resources to target high-need cases and key determinants, which is particularly useful for LMICs and low-resource settings. This paper illustrates an example of how to move toward this direction. Future research is needed to validate the approach.

Despite much recent empirical work on inequity aversion in nonhuman species, many questions remain about its distribution across taxa and the factors that shape its evolution and expression. Past work suggests that domestic dogs (Canis familiaris) and wolves (Canis lupus) are averse to inequitable resource distributions in contexts that call upon some degree of training such as 'give paw' and 'buzzer press' tasks. However, it is unclear whether inequity aversion appears in other canid species and in other experimental contexts. Using a novel inequity aversion task that does not require specific training, this study helps address these gaps by investigating inequity aversion in domestic dogs and a closely related but non-domesticated canid, the dingo (Canis dingo). Subjects were presented with equal and unequal reward distributions and given the opportunity to approach or refuse to approach allocations. Measures of interest were (1) subjects' refusal to approach when getting no food; (2) approach latency; and (3) social referencing. None of these measures differed systematically across the inequity condition and control conditions in either dogs or dingoes. These findings add to the growing literature on inequity aversion in canids, providing data from a new species and a new experimental context. Additionally, they raise questions about the experimental features that must be in place for inequity aversion to appear in canids.

Context It is widely recognised that red foxes (Vulpes vulpes) are abundant within urban areas; however, it is difficult to apply lethal control measures using poison baits in cities because of concerns about the safety of domestic pets, particularly dogs (Canis familiaris). Aims We tested canid pest ejectors (CPEs) as a potential method of fox control by measuring visitation and activation behaviour of foxes and other wildlife while assessing non-target risk to domestic dogs. Methods We compared eight urban and eight peri-urban sites in Sydney, with half of the sites having restricted access for domestic dogs. We allocated five camera traps and ejectors per site. Through generalised linear mixed models, we compared the probability of ejector activation between foxes and dogs. We also assessed the relationship between dog visitation and distance to habitation and dog restrictions as measures of dog safety. Key results Both species of canids were equally likely to pull the ejector (P = 0.26). As expected, dog visitation was significantly lower in sites with dog restrictions (P < 0.001). However, it was not related to distance from habitation. Only two non-canid species were recorded pulling the ejector, suggesting high target-specificity for canids. Conclusions In sites with dog restrictions, the risk of dog casualties from CPEs is minimal. However, distance from habitation does not increase dog safety, at least within 250 m. The ejector is highly specific for canids. Implications We provide specific recommendations for the design of a potential fox control program using CPEs in urban and peri-urban areas. The ejector may be a safe method for fox control in cities when deployed at places without domestic dogs.

The zebrafish (Danio rerio), is an important biomedical model organism used in many disciplines, including development, disease modeling and toxicology, to better understand vertebrate biology. The phenomenon of developmental delay in zebrafish embryos has been widely reported as part of a mutant or treatment-induced phenotype, and accurate characterization of such delays is imperative. Despite this, the only way at present to identify and quantify these delays is through manual observation, which is both time-consuming and subjective. Machine learning approaches in biology are rapidly becoming part of the toolkit used by researchers to address complex questions. In this work, we introduce a machine learning-based classifier that has been trained to detect temporal developmental differences across groups of zebrafish embryos. Our classifier is capable of rapidly analyzing thousands of images, allowing comparisons of developmental temporal rates to be assessed across and between experimental groups of embryos. Finally, as our classifier uses images obtained from a standard live-imaging widefield microscope and camera set-up, we envisage it will be readily accessible to the zebrafish community, and prove to be a valuable resource.

The zebrafish (Danio rerio), is an important biomedical model organism used in many disciplines, including development, disease modeling and toxicology, to better understand vertebrate biology. The phenomenon of developmental delay in zebrafish embryos has been widely reported as part of a mutant or treatment-induced phenotype, and accurate characterization of such delays is imperative. Despite this, the only way at present to identify and quantify these delays is through manual observation, which is both time-consuming and subjective. Machine learning approaches in biology are rapidly becoming part of the toolkit used by researchers to address complex questions. In this work, we introduce a machine learning-based classifier that has been trained to detect temporal developmental differences across groups of zebrafish embryos. Our classifier is capable of rapidly analyzing thousands of images, allowing comparisons of developmental temporal rates to be assessed across and between experimental groups of embryos. Finally, as our classifier uses images obtained from a standard live-imaging widefield microscope and camera set-up, we envisage it will be readily accessible to the zebrafish community, and prove to be a valuable resource.

Endangered San Joaquin kit foxes Vulpes macrotis mutica can be sympatrically distributed with as many as four other canids: red fox, gray fox, coyote and domestic dog. Canid scats are often found during routine fieldwork, but cannot be reliably identified to species. To detect and study the endangered kit fox, we developed mitochondrial DNA markers that can be amplified from small amounts of DNA extracted from scats. We amplified a 412-bp fragment of the mitochondrial cytochrome-b gene from scat samples and digested it with three restriction enzymes. The resulting restriction profiles discriminated among all five canid species and correctly identified 10 'unknown' fox scats to species in blind tests. We have applied our technique to identify canids species for an environmental management study and a conservation study. We envision that our protocol, and similar ones developed for other endangered species will be greatly used for conservation management in the future.

Chapter 5 - Auditory Communication in Domestic Dogs: Vocal Signalling in the Extended Social Environment of a Companion Animal